What's your choice? Learning the mixed multi-nomial logit model

Ammar Ammar, Sewoong Oh, Devavrat Shah, Luis Filipe Voloch

Research output: Contribution to journalConference article

Abstract

Computing a ranking over choices using consumer data gathered from a heterogenous population has become an indispensable module for any modern consumer information system, e.g. Yelp, Netflix, Amazon and app-stores like Google play. In such applications, a ranking or recommendation algorithm needs to extract meaningful information from noisy data accurately and in a scalable manner. A principled approach to resolve this challenge requires a model that connects observations to recommendation decisions and a tractable inference algorithm utilizing this model. To that end, we abstract the preference data generated by consumers as noisy, partial realizations of their innate preferences, i.e. orderings or permutations over choices. Inspired by the seminal works of Samuelson (cf. axiom of revealed preferences) and that of McFadden (cf. discrete choice models for transportation), we model the population's innate preferences as a mixture of the so called Multi-nomial Logit (MMNL) model. Under this model, the recommendation problem boils down to (a) learning the MMNL model from population data, (b) finding am MNL component within the mixture that closely represents the revealed preferences of the consumer at hand, and (c) recommending other choices to her/him that are ranked high according to thus found component. In this work, we address the problem of learning MMNL model from partial preferences. We identify fundamental limitations of any algorithm to learn such a model as well as provide conditions under which, a simple, data-driven (non-parametric) algorithm learns the model effectively. The proposed algorithm has a pleasant similarity to the standard collaborative filtering for scalar (or star) ratings, but in the domain of permutations. This work advances the state-of-art in the domain of learning distribution over permutations (cf. [2]) as well as in the context of learning mixture distributions (cf. [4]).

Original languageEnglish (US)
Pages (from-to)565-566
Number of pages2
JournalPerformance Evaluation Review
Volume42
Issue number1
DOIs
StatePublished - Jun 20 2014
EventACM SIGMETRICS International Conference on Measurement and Modeling of Computer Systems, SIGMETRICS 2014 - Austin, United States
Duration: Jun 16 2014Jun 20 2014

Fingerprint

Collaborative filtering
Application programs
Stars
Information systems

ASJC Scopus subject areas

  • Software
  • Hardware and Architecture
  • Computer Networks and Communications

Cite this

What's your choice? Learning the mixed multi-nomial logit model. / Ammar, Ammar; Oh, Sewoong; Shah, Devavrat; Voloch, Luis Filipe.

In: Performance Evaluation Review, Vol. 42, No. 1, 20.06.2014, p. 565-566.

Research output: Contribution to journalConference article

Ammar, Ammar ; Oh, Sewoong ; Shah, Devavrat ; Voloch, Luis Filipe. / What's your choice? Learning the mixed multi-nomial logit model. In: Performance Evaluation Review. 2014 ; Vol. 42, No. 1. pp. 565-566.
@article{68787506aae24c62b0a88774e31e6d84,
title = "What's your choice? Learning the mixed multi-nomial logit model",
abstract = "Computing a ranking over choices using consumer data gathered from a heterogenous population has become an indispensable module for any modern consumer information system, e.g. Yelp, Netflix, Amazon and app-stores like Google play. In such applications, a ranking or recommendation algorithm needs to extract meaningful information from noisy data accurately and in a scalable manner. A principled approach to resolve this challenge requires a model that connects observations to recommendation decisions and a tractable inference algorithm utilizing this model. To that end, we abstract the preference data generated by consumers as noisy, partial realizations of their innate preferences, i.e. orderings or permutations over choices. Inspired by the seminal works of Samuelson (cf. axiom of revealed preferences) and that of McFadden (cf. discrete choice models for transportation), we model the population's innate preferences as a mixture of the so called Multi-nomial Logit (MMNL) model. Under this model, the recommendation problem boils down to (a) learning the MMNL model from population data, (b) finding am MNL component within the mixture that closely represents the revealed preferences of the consumer at hand, and (c) recommending other choices to her/him that are ranked high according to thus found component. In this work, we address the problem of learning MMNL model from partial preferences. We identify fundamental limitations of any algorithm to learn such a model as well as provide conditions under which, a simple, data-driven (non-parametric) algorithm learns the model effectively. The proposed algorithm has a pleasant similarity to the standard collaborative filtering for scalar (or star) ratings, but in the domain of permutations. This work advances the state-of-art in the domain of learning distribution over permutations (cf. [2]) as well as in the context of learning mixture distributions (cf. [4]).",
author = "Ammar Ammar and Sewoong Oh and Devavrat Shah and Voloch, {Luis Filipe}",
year = "2014",
month = "6",
day = "20",
doi = "10.1145/2637364.2592020",
language = "English (US)",
volume = "42",
pages = "565--566",
journal = "Performance Evaluation Review",
issn = "0163-5999",
publisher = "Association for Computing Machinery (ACM)",
number = "1",

}

TY - JOUR

T1 - What's your choice? Learning the mixed multi-nomial logit model

AU - Ammar, Ammar

AU - Oh, Sewoong

AU - Shah, Devavrat

AU - Voloch, Luis Filipe

PY - 2014/6/20

Y1 - 2014/6/20

N2 - Computing a ranking over choices using consumer data gathered from a heterogenous population has become an indispensable module for any modern consumer information system, e.g. Yelp, Netflix, Amazon and app-stores like Google play. In such applications, a ranking or recommendation algorithm needs to extract meaningful information from noisy data accurately and in a scalable manner. A principled approach to resolve this challenge requires a model that connects observations to recommendation decisions and a tractable inference algorithm utilizing this model. To that end, we abstract the preference data generated by consumers as noisy, partial realizations of their innate preferences, i.e. orderings or permutations over choices. Inspired by the seminal works of Samuelson (cf. axiom of revealed preferences) and that of McFadden (cf. discrete choice models for transportation), we model the population's innate preferences as a mixture of the so called Multi-nomial Logit (MMNL) model. Under this model, the recommendation problem boils down to (a) learning the MMNL model from population data, (b) finding am MNL component within the mixture that closely represents the revealed preferences of the consumer at hand, and (c) recommending other choices to her/him that are ranked high according to thus found component. In this work, we address the problem of learning MMNL model from partial preferences. We identify fundamental limitations of any algorithm to learn such a model as well as provide conditions under which, a simple, data-driven (non-parametric) algorithm learns the model effectively. The proposed algorithm has a pleasant similarity to the standard collaborative filtering for scalar (or star) ratings, but in the domain of permutations. This work advances the state-of-art in the domain of learning distribution over permutations (cf. [2]) as well as in the context of learning mixture distributions (cf. [4]).

AB - Computing a ranking over choices using consumer data gathered from a heterogenous population has become an indispensable module for any modern consumer information system, e.g. Yelp, Netflix, Amazon and app-stores like Google play. In such applications, a ranking or recommendation algorithm needs to extract meaningful information from noisy data accurately and in a scalable manner. A principled approach to resolve this challenge requires a model that connects observations to recommendation decisions and a tractable inference algorithm utilizing this model. To that end, we abstract the preference data generated by consumers as noisy, partial realizations of their innate preferences, i.e. orderings or permutations over choices. Inspired by the seminal works of Samuelson (cf. axiom of revealed preferences) and that of McFadden (cf. discrete choice models for transportation), we model the population's innate preferences as a mixture of the so called Multi-nomial Logit (MMNL) model. Under this model, the recommendation problem boils down to (a) learning the MMNL model from population data, (b) finding am MNL component within the mixture that closely represents the revealed preferences of the consumer at hand, and (c) recommending other choices to her/him that are ranked high according to thus found component. In this work, we address the problem of learning MMNL model from partial preferences. We identify fundamental limitations of any algorithm to learn such a model as well as provide conditions under which, a simple, data-driven (non-parametric) algorithm learns the model effectively. The proposed algorithm has a pleasant similarity to the standard collaborative filtering for scalar (or star) ratings, but in the domain of permutations. This work advances the state-of-art in the domain of learning distribution over permutations (cf. [2]) as well as in the context of learning mixture distributions (cf. [4]).

UR - http://www.scopus.com/inward/record.url?scp=84955573373&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84955573373&partnerID=8YFLogxK

U2 - 10.1145/2637364.2592020

DO - 10.1145/2637364.2592020

M3 - Conference article

AN - SCOPUS:84955573373

VL - 42

SP - 565

EP - 566

JO - Performance Evaluation Review

JF - Performance Evaluation Review

SN - 0163-5999

IS - 1

ER -